Granted Patent B1

US 12,373,935 · App. 19/032,581 · Granted Jul 29, 2025

Inventors: Idan Cohen (Tel Aviv, IL); Ilya Grinshpoun (Ramat Gan, IL); Itai Orr (Or Akiva, IL); Amir Hever (Tenafly, NJ)

Assignee: UVeye Ltd.

G06T7/0004G06T3/4038G06T7/10G06V10/16G06V10/26G06V10/82G06T2200/24G06T2207/20081G06T2207/20084G06T2207/30204G06T2207/30252

Abstract

Claims (89)

1. A system for generating a user interface, comprising:

a set of multiple imaging devices positioned on at least one side of an inspection passage that a vehicle passes through;

at least one processor; and

a memory storing instructions that, when executed by the at least one processor, cause the system to:

obtain, from the set of multiple imaging devices, a plurality of sets of multiple images of a plurality of segments of a surface of the vehicle, the plurality of sets of multiple images are captured at a plurality of time points during a relative movement between the vehicle and the set of multiple imaging devices;

stitch from the plurality of sets of multiple images a dataset record mapping a plurality of vehicle parts and a plurality of anomalies on a surface of one or more of the plurality of vehicle parts;

transform data from the plurality of sets of multiple images to a moving visual media object depicting the vehicle from a plurality of different points of view using a first generative artificial intelligence (AI) model;

compute a mapping record between the plurality of segmented parts of the vehicle and areas in at least one target frame from the moving visual media object using the dataset record;

transform the mapping record and the moving visual media object to an interactive user interface depicting a plurality of user-selectable markers synchronized with a playback of the moving visual media object, indicating locations of the plurality of anomalies on the surface of the one or more vehicle parts from a group of the plurality of different points of view using a second generative AI model;

wherein the interactive user interface performs a data retrieval and display action based on one or more user selection of one or more of the plurality of anomalies.

2. The system of claim 1 , wherein the instructions, when executed by the at least one processor, further cause the system to: sequentially present multiple frames of the moving visual media object visual to present using the interactive user interface to a user a plurality of different points of view in relation to the depicted vehicle.

3. The system of claim 2 , wherein the instructions, when executed by the at least one processor, further cause the system to: analyze the plurality of sets of multiple images using a car parts segmentation model to generate car parts segmentations; perform temporal tracking of the car parts segmentations across the multiple frames of the moving visual media object.

4. The system of claim 2 , wherein the instructions, when executed by the at least one processor, further cause the system to:

identify anchor parts on the vehicle using a car parts segmentation model;

track the identified anchor parts across the multiple frames;

wherein the anchor parts comprise fixed reference points including one or more of: wheels, door handles, headlights, and taillights.

5. The system of claim 4 , wherein the instructions, when executed by the at least one processor, further cause the system to: use the tracked anchor parts to align multiple camera views during the stitching of the dataset record.

6. The system of claim 4 , wherein the instructions, when executed by the at least one processor, further cause the system to: use the tracked anchor parts to register damage locations relative to the anchor parts; and synchronize damage marker positions during a playback of the moving visual media object.

7. The system of claim 1 , wherein the interactive user interface comprises: a main multiple degrees view displaying the moving visual media object; a damage preview panel displaying close-up images and detailed information about detected anomalies; a body part panel providing access to specific vehicle components; an anchor point bar providing navigation to predefined angles around the vehicle.

8. The system of claim 1 , wherein the interactive user interface is configured to: display damage markers overlaid on the vehicle that rotate in sync with the vehicle's movement.

9. The system of claim 1 , wherein the second generative AI model implements:

an attention mechanism to:

focus on relevant vehicle regions during rotation;

predict optimal marker placement;

maintain visual consistency of interface elements; and

adapt visualization parameters based on user interaction patterns.

10. The system of claim 1 , wherein the instructions to compute the mapping record, when executed by the at least one processor, cause the system to:

register the dataset record to a coordinate system of a visual representation;

map a plurality of features from the dataset record to corresponding locations in frames of the visual representation.

11. The system of claim 1 , wherein the instructions, when executed by the at least one processor, further cause the system to:

implement an automated motion function that modifies the visual representation according to predetermined parameters;

maintain a feature overlay system that synchronizes visual indicators with the motion;

execute a visual connection function that generates graphical links between related interface elements in response to user interaction events;

wherein the visual connection function maintains a real-time state record tracking active connections during the motion.

12. The system of claim 11 , wherein the instructions, when executed by the at least one processor, further cause the system to:

automatically trigger motion to predetermined optimal viewpoints stored in a viewpoint database;

execute real-time spatial transformations to maintain visual indicator positioning during motion transitions; and

synchronize the visual connection function with the motion transitions to maintain graphical links between interface elements.

13. The system of claim 1 , wherein the instructions, when executed by the at least one processor, further cause the system to:

maintain a feature registry mapping inspection areas to corresponding detailed views;

implement a display system that dynamically generates dedicated inspection views for selected features;

execute real-time highlighting functions creating visual connections between selected elements and their locations in the visual representation;

wherein the highlighting functions maintain spatial accuracy during motion and view transitions.

14. The system of claim 1 , wherein the instructions, when executed by the at least one processor, further cause the system to:

implement a navigation system comprising a series of predefined viewpoints;

maintain a navigation mapping table correlating navigation elements with specific motion parameters;

execute a motion control function that:

modifies automated motion based on user selection;

performs smooth transition animations between viewpoints while maintaining spatial coherence of visual indicators;

updates indicator positions and visual connections in real-time during transitions;

wherein the motion control function references the mapping record to ensure accurate spatial relationships throughout the transition.

15. The system of claim 1 , wherein the second generative AI model comprises:

a neural network trained to:

generate user interface elements synchronized with vehicle rotation;

predict optimal marker positions across different viewing angles;

maintain temporal consistency of marker positioning during transitions; and

wherein the neural network utilizes anchor part tracking data to validate marker positioning accuracy.

16. The system of claim 1 , wherein the system further implements real-time synchronization by:

maintaining a temporal buffer of marker positions;

predicting marker trajectories during rotation transitions;

interpolating between known positions for smooth animation;

validating marker positions against the mapped anomaly locations; and

adjusting marker visibility based on current viewing angle.

17. The system of claim 1 , wherein the instructions further cause the system to:

implement a damage severity classification model that:

analyzes detected anomalies using multi-scale feature extraction;

assigns severity scores based on size, depth, and location;

generates repair cost estimates; and

updates marker visualization based on severity classification.

18. The system of claim 1 , wherein the first generative AI model implements:

a multi-stage pipeline comprising:

an initial reconstruction stage generating a coarse 3D model;

a refinement stage enhancing geometric detail;

a texture synthesis stage maintaining visual consistency; and

a view-dependent rendering stage optimizing output for current perspective.

19. The system of claim 1 , wherein the system implements marker synchronization by:

maintaining a spatial graph structure linking:

anchor parts positions;

detected anomaly locations;

marker rendering parameters; and

updating the graph in real-time during visualization playback.

20. A method of generating an interactive vehicle inspection interface, comprising:

obtaining, from a set of multiple imaging devices positioned on at least one side of a stationary inspection passage that a vehicle passes through, a plurality of sets of multiple images of a plurality of segments of a surface of the vehicle, the plurality of sets of multiple images are captured at a plurality of time points during a relative movement between the vehicle and the set of multiple imaging devices;

stitching from the plurality of sets of multiple images a dataset record mapping a plurality of vehicle parts and a plurality of anomalies on a surface of one or more of the plurality of vehicle parts;

transforming data from the plurality of sets of multiple images to a moving visual media object depicting the vehicle from a plurality of different points of view using a first generative artificial intelligence (AI) model;

computing a mapping record between the plurality of segmented parts of the vehicle and areas in at least one target frame from the moving visual media object using the dataset record;

transforming the mapping record and the moving visual media object to an interactive user interface depicting a plurality of user-selectable markers synchronized with a playback of the moving visual media object, indicating locations of the plurality of anomalies on the surface of the one or more vehicle parts from a group of the plurality of different points of view using a second generative AI model;

controlling a presentation of the moving visual media object by the interactive user interface in response to a user action indicative of a selection of one or more of the plurality of user-selectable markers by a user;

wherein the interactive user interface performs a data retrieval and display action based on one or more user selection of one or more of the plurality of anomalies.

Assignments (1)

References Cited (9)